Re: [PATCH] timing.cc integer overflow

[David Fifield <david () bamsoftware com>]

On Mon, Feb 09, 2009 at 08:19:06PM +0100, Henri Doreau wrote:
> Thanks, I have committed this patch. However I think there's still some
> > work to do. The member function ScanProgressMeter::printStatsIfNecessary
> > does some of the same calculations as ScanProgressMeter::printStats, and
> > it still does them with longs. I don't know what the effect would be,
> > but it could cause the stats not to be printed when they should be, or
> > vice versa. Can you look at that function and do the calculations in the
> > same way, or (even better) remove the duplicated code into a separate
> > function?
>
> I have added a (public) method: ScanProgressMeter::estimateTimeLeft that
> returns a double. It "factorizes" the duplicated code for both
> ScanProgressMeter::printStatsIfNecessary and ScanProgressMeter::printStats.
> So the patch contains a modified version of
> ScanProgressMeter::printStatsIfNecessary, and the same way,
> ScanProgressMeter::mayBePrinted was changed to compute things in seconds
> only.

I applied your patch with minor changes. I made estimateTimeLeft into a
static function within timing.cc. The little trick where 0.5 is added to
the remaining time to get the effect of rounding in integer
calculations, I moved out of estimateTimeLeft, to just before the
integer calculations are done.

David Fifield

_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://SecLists.Org